3 Kordofanian

Thilo C. Schadeberg

Kordofanian languages are spoken in the Nuba Mountains in the Republic of the Sudan. The name 'Kordofanian' is an unfortunate choice since in the Sudan, particularly in the Nuba Mountains, the geographical and political term 'Kordofan' refers to the area north of the Nuba Mountains around El Obeid. It might be désirable to coin a more suitable and locally more acceptable name for this language family.

The present survey chapter is based on literature and data col- lected by myself in 1974-75.

3.1 EARLY SCHOLARSHIP

In 1838 Duke Maximilian in Bavaria bought the freedom of four men on the slave market in Alexandria. Taking them home, he put them in the care and under the instruction of his former tutor, Karl Tutscheck, who feit a language could be studied with the help of a single speaker even at gréât distance from that person's linguistic community. One of the four young Africans was Djalo Djondan Are from Tumale, and part of the information he provided on his people and language was published between 1847 and 1853 by Karl's brother Lorenz Tutschek. Prior to this, only two short Kordofanian word lists had been published, one on 'Takele' and one on 'Schabun' (Rüppell 1829). Nothing eise followed until the end of the 19th Century except another small word list from 'Tegelé' (Munzinger 1864). A füll description of this language by Munzinger came into the possession of Reinisch at Vienna but bas since been destroyed. The fact that the Tegali dialect cluster figures in each of the early linguistic con- tacts with Kordofanian might be explained by the relative importance of the Tegali kingdom in the 19th Century.

The signifïcance of this area for African language classification was realized by Lepsius (1880). In this Century several linguistic surveys have been carried out in the Nuba Mountains (Seligman 1910/11; Meinhof 1915-1919; MacDiarmid and MacDiarmid 1931; Stevenson 1956-1957; Schadeberg 1981a, b). Gradually, the extent of the Kordofanian language 68 Schadeberg family, its grammatical and lexical characteristics, and ils internai relationships have become apparent.

Nevertheless, the quality of this survey chapter is severely limited by the scarcity of in-depth descriptions. Meinhof (1943/44) wrote the first short grammar of any Kordofanian language, a sketch of Ebang (Heiban) based on analysis of published translations of two Gospels. The most complete existing description of any Kordofanian language is a grammar of Moro by Black and Black (1971). Fr. Carlo Muratori's notes on Orig, another dialect of the Tegali cluster, were reanalyzed and edited by Schadeberg and Elias (1979). As for the Kadugli language group (not hère considered to be Kordofanian), Reh (1985) provides an extensive and detailed description of Krongo.

3.2 CLASSIFICATION

3.2.1 INTERNAL CLASSIFICATION AND LANGUAGE INVENTORY

There is one level of classification within Kordofanian that is unambiguous and non-controversial. I refer to units at this level as 'groups', and there are four such groups:

1. HEIBAN (Koalib; Koalib-Moro) 2. TALODI (Talodi; Talodi-Masakin and Lafofa) 3. RASHAD (Tegali; Tegali-Tagoi) 4. KATLA (Katla; Katla)

I have added in parentheses the group names first as used by Greenberg (1963) and then those of Tucker and Bryan (1956). The group names I propose are the names of towns centrally located in the respec- tive areas; the riverine terminology populär in African linguistics cannot be applied to the Nuba Mountains for lack of permanent surface water. No confusion should arise between group names and language names since none of the place names are identical with the name any particular linguistic grouping gives itself.

The overall division into groups was first established by MacDiarmid and MacDiarmid (1931). The only amendment to their grouping was Greenberg's (1950) inclusion of Tegem (Lafofa) in the Talodi group.

There is a fifth group which has been classified as Kordofanian by Greenberg since 1950, i.e., KADUGLI (Tumtum; Kadugli-Krongo). I propose to exclude Kadugli from Kordofanian (see sec. 3.2.3). • HEIBAN-Group • TALOOI - Group T RASHAD-Group • KATLA - Group * KAOUGLI-Group

The distribution of Kordofanian tanguages • c f c »•««»•'• »«o (Sou t h «f n Kordofon Provinc«. Sudan)

PLATE 3 Kordofanian languages 70 Schadeberg

TALODI group: (speakers)

Ngile (Daloka, Masakin) l 12,000 Dengebu (Reikha, Masakin) -' Tocho (Moro Hills settlements) 3,800 Jomang (Talodi) 1,500 Nding (Eliri) 500 Tegem (Lafofa) 3,000

20,800

20% 40% 60% NT 80% | 1=^\ \ Ngile Dengebu Tocho Jomang Nding Tegem

Tegem is very different from the other languages of the Talodi group but there is good lexical, phonological, and morphological évidence for joining Tegem and 'Narrow Talodi' into one group or branch of Kordofanian. I cannot exclude the possibility that some alleged dialects of Masakin and of the cluster called 'Moro Hills' may turn out to be sepa- rate languages.

RASHAD group: (speakers)

Noun dass dialects (Tagoi cluster): Goy (Tagoi), Orig (Turjok), Umale (Tumale), etc. 11,000 Classless dialects (Tegali cluster): Gom (Rashad), Tegali, etc. 25,000

36,000

A breakdown into dialects with and without noun classes appears to be thé basic genetic division of the Rashad group. The two dialect clusters share about 70% basic vocabulary (100-word list). There are a number of outlying hills about which no linguistic data is available, but where closely related dialects are supposed to be spoken.

KATLA group: (speakers)

Kalak (Katla) 10,900i Lomorik (Tima) 1,100?

12,000? Kordofanian 7l

The two Katla languages share about 50% basic vocabulary (100- word list). Even less is known about this group than about the others.

It is not clear at present whether thé four groups should be re- garded as primary branches of Kordofanian or whether intermediate lev- els of relationship should be recognized. Tucker and Bryan originally (1956) posited a 'Larger Unit' consisting of the three groups Heiban, Talodi (and their 'Single Unit' Lafofa), and Rashad. Later (1966), they separated Rashad from this 'Larger Unit' and also drew attention to Stevenson's (1957:51) observation that spécifie lexical resemblances exist between Rashad and Katla.

At least since Meinhof (1915-19) thé question of historical sub- grouping and intergroup relationships has been complicated by confusing it with typological classification into languages with and without noun classes.

3.2.2 EXTERNAL CLASSIFICATION WITHIN NIGER-CONGO

One of Greenberg's 1963 revisions of his own earlier classification of African languages was to join Niger-Congo and Kordofanian äs two primary branches of a single family he called Niger-Kordofanian (also: Congo-Kordofanian). Since then, the genetic relatedness between Kordofanian and Niger-Congo has not been seriously questioned. Still, some linguists believe this kind of historical relation to be too remote to be recoverable by conventional methods of historical linguistics (e.g., Möhlig 1983:175). Convincing lexical resemblances between Kordofanian and other branches of Niger-Congo are scarce. Though Greenberg (1963) lists forty-five sets of possible cognâtes (not counting seven sets linking Niger-Congo exclusively with Kadugli), it remains to be seen which and (less importantly) how many of these will survive the test of internai Kordofanian and Niger-Congo reconstruction. Bennett and Sterk (1977) concluded that their lexical évidence was insufficient for proving genetic relationship between Niger-Congo and Kordofanian (and Mande). Not wishing to reject such links altogether, they assumed an initial three-way split; this position is maintained in the present volume (Niger-Congo Overview).

While the extent and importance of lexical évidence is still un- certain, there is other évidence for which the only possible explanation seems to be common historical origin: similar noun dass Systems appear in most major groupings of Niger-Congo and Kordofanian. At first sight, this may appear to be no more than a typological similarity, and as such, irrelevant to the question of genetic relatedness. However, it has to be 72 Schadeberg pointed out that this type of noun dass System is, if not unique, extremely rare in the languages of the world; the noun dass Systems found in Caucasian and Australian languages are quite different. It is also rare for a language to acquire a noun dass System through borrowing, Mbugu being the only such case ever cited. It is therefore justified to view the présence of this particular type of noun class System as an argument for genetic relations within Niger-Congo, as indeed Greenberg did (1950:393): "The very fact of the existence of noun affix classes for singular and plural in Niger-Congo and Kordofanian is indicative of possible further con- nections, since such formations are rare elsewhere in thé world." At this time, Greenberg did not yet claim Niger-Congo and Kordofanian to be related. Later (1972:190) hè wrote: "It is not too much to state that in at least one instance, that of Niger-Kordofanian, grammatical resemblances were treated as more décisive than lexical ones." However, similarities between the Niger-Congo and Kordofanian noun class Systems are not only typological but can be extended to proper sound-meaning correspondences as well. The meaning of gender is to be understood not only as the vague common semantic feature of part of the nouns occurring in it, but also as that set of nouns (i.e., their meanings) itself. In the figure below (adapted from Schadeberg 1981e), six noun classes of Kordofanian are compared with what I believe are cognate forms in Atlantic, Kwa, Benue-Congo, Gur, and Adamawa-Ubangi.

1 3 4 5 6 6a plof 3 plof 5

Kordofanian gu- gu- j- H- gu- B- people tree egg liquids Atlantic go- gu- Ci- de- ga- ma- (Doneux 1975) people tree names head, name liquids Oti-Volta n -bo -Ci -djt -a -ma (Manessy 1975) people tree egg, head liquids Togo Remnant 0- o- i- U- a- N- (Heine 1968) people firewood egg, head, name liquids Benue-Congo u- u- (t)i- li- a- ma- (De Wolf 1971) people tree egg, head, name liquids Bantu NP mu- mu- mi- j- ma- nia- (Meeussen PP ju- gu- gi- li- ga- ga- 1967) people tree egg, name liquids Mba (Ubangi) -V -V -e -le -me (Carrington Num g- w- 0- 1- 1949) man name liquids

FIGURE 3.1. Comparison of six Kordofanian noun classes Kordofanian 73

The exact place of Kordofanian within Niger-Congo is less clear than its général relation to it. Regarding Kordofanian as one of the three primary branches of Niger-Congo is the least spéculative guess possible in thé absence of spécifie investigations. Future research may show Welmers' hunch to be correct (1973:17): "It is possible that thé relationship between thé Kordofanian languages and the non-Mande Niger-Congo languages is closer than thé relationship of any of them to Mande."

3.2.3 THE KADUGLI GROUP

The Kadugli group comprises nine rather closely related languages. Below I give their names as I recorded them, except in thé case of Krongo, where a phrase meaning 'home-speech' is actually used. In parentheses I add the name of thé village from which I obtained my data, plus thé names of some other villages where the same language is spoken. These village names are more commonly used to identify languages and dialects than thé proper names listed hère. The numbers of speakers were derived in thé same way as for thé Kordofanian languages earlier.

KADUGLI group: (speakers)

Yega (Keiga-Timero, Keiga-al-Kheil, Demik) 7,500 Mudo (Kamdang, Tulishi) 11,000 Talla (Kadugli) 18,000 Miri (Miri Juwa, Miri Barra) 14,000 Tolubi (Tuna, Katcha) 7,000 Kufo (Kufa, Kanga) 8,000 Sangali (Tumma, Krongo-Abdallah) 6,500 Krongo (Tabanya, Dimodongo) 17,000 Talasa (Talasa, Kurondi, Tumtum) 1,600

90,600

60% 70% w K ClK E K 1 80% l i 90% Yega Mudo Talla Miri Tolubi Kufo Sangali Krongo Talasa

The above tree was produced by a preliminary lexicostatistical subclassification that, impressionistically, appears to be supportable by morphological évidence. For example, Central Kadugli distinguishes three genders by concord; Eastern Kadugli has an additional plural 'gender'; only Yega (Western Kadugli) has no such concord (see Stevenson 1957:55-56). 74 Schadeberg

Right from the start Greenberg noted that Kadugli "shows considérable divergence from the remainder" (1963:149). Goodman, in a review of Greenberg (1963), goes one step further: "Particularly dubious is the relationship of the Kadugli-Krongo group to the remaining Kordo- fanian languages" (1970:121). I maintain that it has not been shown that Kadugli is part of Kordofanian, nor that it should be classified as Niger- Congo (see Schadeberg 1981 d).

33 PHONOLOGICAL CHARACTERISTICS

The phonetic and phonological characteristics of Kordofanian languages are poorly documented and analyzed. The most detailed description (Black and Black 1971) concerns Moro (Heiban group which has an atypically complex system of consonants and consonant clusters. This section présents what I take to be two typical sets of phonological proporties of Kordofanian languages. Many of these features may have an areal distribution as well. It remains to be seen what their significance for Niger-Congo comparative studies might be.

3.3.1 POINTS OF ARTICULATION (CONSONANTS)

Most Kordofanian languages distinguish five places of articulation for oral stops: bilabial, laminai (inter)dental, apical alveolar or retroflex, palatal, and velar (eg., /b, d, d/d^ j, g/).

Thus there are two distinct places of articulation in thé dental- alveolar région, and thé distinction is enhanced by an additional différ- ence in the shape of the longue. The dental articulation appears to be more fréquent and less marked than thé alveolar or retroflex one. Speakers of languages lacking this opposition generally use thé dental articulation. Such languages are the two dialect clusters of the Rashad group, but also non-Kordofanian languages such as For (of Darfur) and Sudanese Colloquial Arabie. (I have described the areal spread of this feature in Schadeberg (forthcoming).)

Only four points of articulation are contrastive for nasals. Generally, these are bilabial, alveolar, palatal, and velar. However, I hâve recorded dental nasals in two dialects of Masakin (i.e., in Dengebu) and, as an idiolectic variant, in Ngile. In each case thé dental nasal takes thé place of and corresponds to thé palatal one. Kordofanian 75

m l m

The presumed change from /ji/ to /n/ does not entail a merger with /n/ since it probably arises from a super-distributed articulation in which the contact between the tongue and the roof of the mouth extends from the palatal to the dental région.

Double and secondary articulations are not typical for Kordo- fanian languages. Labial-velars /kp/ and /gb/ occur only in Kalak (Katla group). Labialization is found in some languages of the Heiban and Talodi groups. In Jomang, labialization of bilabial and velar consonants appears to be conditioned by a preceding rounded vowel, e.g., daa- joomwai) 'Jomang village'.

3.3.2 MANNER OF ARTICULATION (OBSTRUENTS)

A striking feature of Kordofanian consonant Systems is the low functional load of the voiced/voiceless distinction. Obstruents are generally voiced between sonorants (vowels, nasals, liquids) and voiceless in obstruent séquences as well as word-finally. Word-initially, much fluctua- tion between voiced and voiceless obstruents is found both within and across languages. For particular languages it may be necessary to recog- nize an unconditioned voicing contrast, but I have not found it necessary to do so in reconstructing proto-Heiban (PH), proto-Talodi (PT), and proto-Rashad (PR) lexical items. A possible source for synchronie intervocalic voiceless stops is thé simplification of long stops, which I analyze as séquences of identical stops.

Intervocalic lenition of obstruents commonly changes voiced stops into fricatives - except for d/

PP H tt ce kk p | t c k b d 4 j g P » t y y Voiceless fricatives are generally rare. A labiodental /f/ as distinct from /p/b/ occurs in the Rashad group. The alveolar fricative /s/ bas a wider distribution throughout thé four branches of Kordofanian. I hâve reconstructed *s for PR corresponding to *ts in PT and to /c/j/s/ in Kalak. Thus it appears that proto-Kordofanian (PK) had an affricate *ts from 76 Schadeberg which many instances of present-day /s/ are derived. A further step is the weakening of /s/ to /h/, as in Goy (Rashad group) and Lomorik (Katla group).

3.4 MORPHOLOGICAL CHARACTERISTICS: NOUN CLASSES

Three of the four Kordofanian language groups (Heiban, Talodi, and Rashad) hâve noun class Systems strongly reminiscent of those found elsewhere in Niger-Congo. I hâve attempted to reconstruct noun class Systems for thé three groups. (From hère on, d and t (without any dia- critic) represent dental obstruents; voicing is not distinctive in reconstructions.)

, HEIBAN TALODI RASHAD sing. plur. sing. plur. sing. plur.

'gu-'x— *li- *b-^X— *(y)- {LAB}N 'g- ^^*j- *w- — >*y- >*g- *w—— ^*n- •j--^— *m- *d-<^ *li- ^— *nu- *d--^— *r- 'y--^^'g- *3 ^*ji- *<- —^^*n- »g -*s- #r4 *H " q. — *w- *jt -*J1- *ts ^•J«-

FIGURE 3.2. Reconstructed noun class Systems

The two languages of thé fourth group, Katla, also have number marking by préfixes, but thèse are less numerous and there is no concord.

Reconstruction of the PR noun class System relies totally on évi- dence from thé Tagoy dialects. Différent explanations exist for why thèse dialects hâve a fully operative noun class System that thé Tegali dialects lack. Meinhof (1916:71-72) feit the two languages could not possibly be related genetically and that their close lexical resemblances must, therefore, be due to extensive borrowing. Greenberg (1963), who excludes the possibility of mass borrowing of basic vocabulary, regards the two dialect clusters as closely related and assumes the loss of noun classes in the Tegali dialects. Tucker and Bryan (1966:270) agrée that the two dialect clusters must be closely related, but since at that time they were no longer convinced that Rashad was part of a bigger genetic unity (i.e., Kordofanian) they explain the resemblances between the noun class systems of Tagoy on the one hand and of Heiban and Talodi on the other by Kordofanian 77

'adoption', i.e., Tagoy is supposed to have borrowed its whole noun dass System. I find this hypothesis unlikely.

Firstly, such borrowing of a whole noun dass System - if it exists at all (cf. Mbugu!) - is extremely rare. Secondly, the Tagoy noun dass System is characteristically different from any System found in the Heiban or Talodi groups; in other words, we know of no possible donor language. Thirdly, the borrowing hypothesis entails a number of highly spécifie historical assumptions for which no independent évidence is available: PK must have had a noun dass System since three of its descendants did preserve it: Heiban, Talodi, and the unknown language (X) from which Tagoy borrowed it; Rashad supposedly lost it; then Tagoy borrowed it from X; and finally X disappeared.

I prefer to follow Greenberg (1963) and simply assume the loss of noun classes in Tegali. This hypothesis is supported by what I take to be remuants of earlier noun class préfixes in Tegali, e.g., a great number of place names starting with t- and some instances of a prefix g-, which is used to form abstract nouns and language names. For instance, the Tegali dialect spoken in the town of Rashad is called nagom, and the town itself is called togont.

The figure 3.3 is an attempt to line up possibly related noun class préfixes from Heiban, Talodi, and Rashad, together with approximations towards proto-Kordofanian reconstructions.

Noun classes l and 3 through 6a have been used above to show that Kordofanian is part of Niger-Congo. The correspondences appear to be well established.

Gender 1/2 is the 'personal' gender. It regularly contains eth- nonyms and, in Heiban and Talodi,'also the words 'person/people', 'man', and 'woman'. Some animais also take this gender, e.g., 'bird', 'dog', 'fish', 'snake', but there is usually an alternative gender for these nouns. (Compare Rere ku-jaw/la-faw 'big bird' with ta-r,aw/jia-r.aw 'small bird'.) The sound correspondence PH *gu- to PT *b- has its exact nasal parallel in the corresponding préfixes of class 6 (PH *gu- : PT *m-). I assume that PT *y- is a regulär reflex of PK *li-. No word-initial *1 has been reconstructed for PT, but there are a few compounds in Tegem (Lafofa) where *1 appears to have been preserved word-internally, e.g., im-b-ügi/i- 1-cgi 'my wife/my^wives'. 78 Schadeberg

HEIBAN TALODI RASHAD KORDOFANIAN 1. sg. *gu- *b- *w- *gu- 2. pi. *li- *y- *y- *li- 3. sg. *gu- *b- *w- *gu- 4. pi. •j- *g- *y- *gi- 5. sg. *li- •j- *y- *li-7 6. pi. *gu- *m- *5- *OU- 6a. coll. *g- *0- *o- *o-

7. sg. *ts- *j- •j- 8. pi. *J»- 'Ji- *Ji- •j>- 9. sg. *o- *o- *5- 10. pi. •j- *s- *s- 11. sg. *g- *g- *g- *g- 12. pi. *n- *n- *n- 13. sg. *d- *d- *d- *d- 14. pi. *4- *r- •«1-7 15. sg. *4- *4- *d- *4- 16. ? *w- {LAB} ?

FIGURE 3.3. Noun class préfixes

Gender 3/4 contains the word for 'tree' in Heiban and Talodi, and a whole series of spécifie tree names in Rashad. Class 3 is the same as class 1; it bas been assigned its own number only for casier référence in the wider context of Niger-Congo comparison. In Rashad, classes 2 and 4 hâve merged.

The most stable item in gender 5/6 is 'egg'; Heiban and Talodi also attest 'head', 'bone', and 'star'. It is not clear what phonological dif- férence should be reconstructed to account for thé distinct Talodi réflexes *y- and *j- of PK classes 2 and 5 *li-.

The one-class gender 6a is for liquids and contains words such as 'blood', 'fat', and 'water'. The phonological différence with thé class 6 prefix is minimal: Heiban *flu- and *n- are generally not distinct before rounded vowels. The two class préfixes hâve completely merged in Rashad, but they are nearly distinct in Talodi.

Evidence for thé other classes is less compelling if we insist on more than phonetic similarity. Gender 13/14 might vaguely be character- ized as thé gender of 'elongated things'; likely reconstructions include items such as 'horn', 'root', 'tau', and probably also 'tongue'. Rashad bas no direct réflexes of this gender, which is not surprising since Rashad Kordofanian 79 does not distinguish PK *4. from PK *d and bas replaced thé plural dass 14 by the existing plural dass 6.

3.5 CONCLUSION

Descriptive and comparative work on Kordofanian languages is still in its infancy. The study of Kordofanian deserves high priority in view of the isolated but prominent position this linguistic unit bas within Niger- Congo. The linguistic distance between thé four groups is considérable. Kordofanian has been presented hère as related to Niger-Congo, but so far no particular links with any subgroup of that language family have been discovered. Available évidence points to a long présence of Kordofanian in its présent location, thé Nuba Mountains, which it shares with several other language groups. Except for Kadugli, these have all been classified as Eastern Sudanic. It has been argued (Thelwall and Schadeberg 1983) that of ail those groups, Kordofanian represents thé oldest linguistic layer in thé Nuba Mountains. This will be an important point to consider when the time cornes to address thé question of the origin of Niger-Congo.

BIBLIOGRAPHY

Bennett, Patrick R. and Jan P. Sterk. 1977. South Central Niger-Congo: A reclassification. SAL 8:241-73. Black, K. and Mrs. K. Black. 1971. The Moro language: Grammar and dictionaiy. Linguistic Monograph Séries 6. Khartoum: Sudan Research Unit, Faculty of Arts. Carrington, J.F. 1949. Esquisse de la langue mba. Kongo-Overzee 15:90-107. De Wolf, Paul P. 1971. The neun dass System ofProto-Benue-Congo. The Hague: Mouton. Doneux, J.L. 1975. Hypothèses pour la comparative des langues atlantiques. Africana Linguistica 6:41-129. Goodman, Morris. 1970. Some questions on thé classification of African languages. UAL 36:117-22. Greenberg, Joseph H. 1950. Studies in African linguistic classification: 7. Smaller families; Index of languages. Southwestem Journal of Anthropology 6:388-98. . 1963. The languages ofAfrica. The Hague: Mouton; UAL 29.1, Publication 25 of thé Indiana University Research Center in Anthropology, Folklore and Linguistics. Bloomington: Indiana University Press. 1972. Linguistic évidence regarding Bantu origins. JAfrH 13.2:189-216. Heine, Bernd. 1968. Die Verbreitung und Gliederung der Togorestsprachen. Kölner Beiträge zur Afrikanistik 1. Berlin: Dietrich Reimer Verlag. Lepsius, Richard. 1880. Nubische Grammatik mit einer Einleitung über die Völker und Sprachen Afrika's. Berlin: Wilhelm Hertz. MacDiarmid, P.A. and D.N. MacDiarmid. 1931. The languages of the Nuba Mountains. Sudan notes and records 14:149-62. 80 Schadeberg

Manessy, Gabriel. 1975. Les langues oti-volta. LACITO 5. Paris: SELAF. Meeussen, A.E. 1967. Bantu grammatical reconstructions. Africana linguistica 3, Annalen van het Koninklijk Museum voor Midden-Afrika 61:79-121. Tetvuren. Meinhof, Carl. 1915-19. Sprachstudien im egyptischen Sudan. Zeitschrift für Kolonialsprachen 6(1915/16):161-205, 264-84; 7(1916/17):36-80, 110-33, 212-50, 326-35; 8(1917/18):46- 74, 110-39, 170-96, 257-67; 9(1918/19):43-64, 89-117, 167-204, 226-55. . 1943/44. Das Heiban in Kordofan. Zeitschrift ßr Eingeborenen-Sprachen 34:94- 130. Möhlig, Wilhelm J.G. 1983. Niger-Kordofanisch. Lexikon der Afrikanistik, ed. by H. Jungraithmayr and W.J.G. Möhlig, pp.175-76. Berlin: Dietrich Reimer. Munzinger, Werner. 1864. Werner Munzinger's Bericht über seine Reise von Massua nach Kordofan, 1861 und 1862. Die deutsche Expedition in Ost-Afrika, 1861 und 1862, ed. T. v. Heuglin et al., pp. 1-13. Ergänzungsheft No. 13 zu Permann's 'Geographischen Mittheilungen'. Gotha: Justus Perthes. Reh, Mechthild. 1985. Die Krongo-Sprache (niino mo-di): Beschreibung, Texte, Wörterver- zeichnis. Kölner Beiträge zur Afrikanistik 12. Berlin: Dietrich Reimer. Rüppell, Eduard. 1829. Reisen in Nubien, Kordofan und dem peträischen Arabien. Frankfurt am Main: Friedrich Wilmans. Schadeberg, Thilo C. 1981a. A survey of Kordofanian. Vol. I: The Heiban group. SUGIA Beiheft 1. Hamburg: Helmut Buske Verlag. . 1981b. A survey of Kordofanian. Vol. II: The Talodi group. SUGIA Beiheft 2. Hamburg: Helmut Buske Verlag. . 1981c. Das Kordofanische. Die Sprachen Afrikas. Band 1: Niger-Kordofanisch, ed. by B. Heine, T.C. Schadeberg, and E. Wolff, pp. 117-28. Hamburg: Helmut Buske Verlag. . 1981d. The classification of the Kadugli language group. Nilo-Saharan, ed. by T.C. Schadeberg and M.L. Bender, pp. 291-305. Dordrecht: Foris Publications. . forthcoming. Zwei areale Sprachmerkmale im Ostsudan. Vienna. and Philip Elias. 1979. A description of the Orig language (Southern Kordofan) based on the notes of Fr. Carlo Muratori. Archives d'anthropologie 26. Tervuren: Musée Royal de l'Afrique Centrale. Seligman(n), Brenda Z. 1910/11. Note on thé language of the Nubas of Southern Kordofan. Zeitschrift für Kolonialsprachen 1:167-88. Stevenson, Roland C. 1956-57. A survey of thé phonetics and grammatical structure of the Nuba Mountain languages, with particular référence to Otoro, Katcha and Nyimang.^4uf/40(1956):73-84, 93-115; 41(1957):27-65, 117-52, 171-96. Thelwall, Robin, and Thilo C. Schadeberg. 1983. The linguistic seulement of the Nuba Mountains. SUGIA 5:219-31. Tucker, Archibald Norman. 1966. Linguistic analyses: The non-Bantu languages of north-eastem Africa. Handbook ofAfrican languages. London: Oxford University Press for IAI. and Margaret A. Btyan. 1956. The non-Bantu languages of north-eastem Africa. Handbook ofAfrican languages, Part III. London: Oxford University Press for IAI. Tutschek, Lorenz. 1848. Über die Tumale-Sprache. Gelehrte Anzeigen, hrsg. von Mitgliedern der k. bayer. Akademie der Wissenschaften, nos. 91-93: columns 729-52. Bulletin der königl. Akademie der Wissenschaften 29-31. . 1848-50. On the Tumali language. Proceedings of the Philological Society for 1846- 47 and 1847-48 3:239-54. PPSfor 1848-49 and 1849-50 4:138-39. Welmers, William E. 1973. African language structures. Berkeley: University of California Press.